Knowledge Work Is Already Over in America

Every job that takes data in, processes it, analyzes it and hands back a recommendation, which is to say knowledge work as a whole, could be replaced by AI within 2 to 5 years. That warning was where the evening started.

This piece is based on the talk Jeewon Kim Serrato gave at a Bloom event.

Why a lawyer of 21 years chose to start a company at 47

The speaker that night was Jeewon Kim Serrato, a Korean American lawyer who lives in San Francisco. She emigrated at 12, finished her undergraduate degree and her law degree at Berkeley, and has spent 21 years in the United States working on data privacy and digital transformation. She led the data privacy practice at the Department of Homeland Security for 8 years, was chief privacy officer at a financial institution holding $4 trillion in assets, and has advised Meta, eBay, LVMH, Sony, Citibank and nearly every other Fortune 500 company on data strategy and digital transformation. She introduced herself as the most general specialist in the world, someone with very narrow expertise held in a very broad and intersecting way.

In February this year she left her partnership at a large law firm, and the reason was simple. She had reached 47 without ever having done zero to one. She had taken the adviser's chair thousands of times and had never once been the founder herself. She now builds and runs an AI-native financial services startup for the American middle class, out of San Francisco. That switch is the backdrop to everything she said, and it gave her account the weight of testimony from someone watching the change facing knowledge workers from the furthest forward seat in the room.

The war on knowledge workers

Her company began from a single scene. In the spring of last year she looked at legal tech AI built to help lawyers and knew straight away that every lawyer would lose their job and that her own was not safe either. In the same stretch four of her friends lost their jobs, eBay let around 600 people go and Intuit around 1,000 people, while Meta and Google carried on cutting. Some of that uses AI as a convenient excuse, she said, but either way the people are out the door.

She cited the occupational research from OpenAI and Anthropic and said the top 40 to 60 occupations can already be performed by AI. Accountants, lawyers, even doctors, every job that takes in data, processes and analyzes it and returns a recommendation. She counts herself among the best lawyers of a 20 year career, and said even she cannot compete with an AI that draws on the knowledge of the whole world, and that within 2 to 5 years knowledge workers will not be needed.

Her prescription came in two branches. Keep pivoting and adapting, or become the owner of your own company and your own path. The choice she put to the room was whether to stay on a salary building somebody else's dream, or build a dream of your own that is distinctive enough for people to believe in.

Silicon Valley as the canary in the coal mine

She calls herself a Silicon Valley insider. Most of her Berkeley classmates are founders or VCs, and she has been working on AI since she was advising self-driving car projects 15 years ago. AI is not new, she said, and what has changed is that everybody uses it now, rather than only the largest companies in the world. That is why she described Silicon Valley as a canary in the coal mine, sensing the change a few years before the rest of the world does.

The cultural gap was the interesting part. Silicon Valley right now is thick with AI pessimism and the United States looks at AI more skeptically than anywhere, while Korea takes it more positively than any country in the world. Her read on where the money sits was just as cold. 80 percent of global venture capital is in Silicon Valley, and 80 percent of that money goes back to Silicon Valley founders, so with a proper founder story and a proper idea an investment offer will arrive off a sketch on a napkin.

She also pointed out that this bubble is bigger and faster than the dot com bubble was. And AI technology has already become political. The United States is a country of 50 states, federal and state AI policy differ completely, and privacy and AI regulation are led by blue states such as California, Washington and New York.

The timing for the United States, and a friend on the ground

Asked about the timing for entering the United States and raising money, the answer was firm. Start now, today. The window may close within 1 to 2 years and has already started closing, so get inside a community and contact every person you can reach, down to the distant aunt or cousin, and build connections out of them. Her view of market size was just as plain. California on its own is bigger than Korea, and the United States has 50 of those, so do not look only at the Korean market.

What she stressed hardest was the local guide. She told a small story about a hotel gym in Busan that required indoor slippers, which she had not known despite coming to Korea every year, and said the reverse happens daily to a Korean founder who lands in Silicon Valley. A co-founder or a local friend who understands the culture is the real condition for success. She also said Silicon Valley is too crowded for fundraising and that other hubs deserve a look.

The alternative she raised was New Mexico. Oil has given it the second largest sovereign wealth fund in the United States, it holds three national laboratories and the highest proportion of PhDs in the country, and it is deep tech ground. 90 percent of Silicon Valley money goes to Stanford MBA founders building B2B SaaS, so a team working on robotics, physical AI, energy infrastructure or logistics is a better match for Albuquerque.

Homogeneity as a weapon rather than a weakness

Korean homogeneity, she said, cuts both ways. She has been told that when a lot of people decide together they stay on safe bets and never take enough risk. She wanted to propose the exact opposite, which is to turn that collective thinking and sense of community into a strength. Unless you are going to move to Silicon Valley, she thought it was better to build a community that grows together where you already stand. She met 15 young founders in Busan and suggested that rather than each of them starting a company, the 15 of them build one, and then ran a brainstorm on how their different capabilities would combine.

She also picked at the difference in how people communicate. Americans are direct, so a pitch deck opens with the conclusion, which is why we win, because backgrounds vary so widely that nothing lands unless you say it outright. A homogeneous population carries context that does not have to be spoken, so the conclusion gets pushed to the back. The practical implication is that conclusion-first American communication has to be practiced on purpose for a global stage. The name she gave her own venture, The K University, came out of asking what she could do differently as a Korean American rather than as a privacy expert.

Breakthrough, and the business of selling happiness

The core concept of The K University is breakthrough. She introduced Pattern Breakers, the book American VCs have read in common over the past 2 years and adopted as a shared language, and set out two conditions for a breakthrough. One is the inflection point, the enormous change happening outside you, such as AI, social media or geopolitics. The other is the insight that belongs to the founder alone and catches that change. You need the ability to observe that the world is moving and your own insight to grab it, both at once. What frustrated her were founders who take the easy road instead of the thing only they can put on the table.

She drew a cold line between a unicorn and a lifestyle business. A VC will not invest in a lifestyle business earning 3 million won a month, and a unicorn is a company that can grow to a billion dollars across 100 countries. If a company cannot get that big it is natural that a VC does not pick up the phone, so rather than reading the silence as failure you should understand that you are simply not the target.

Then she left the heaviest conclusion of the night. What she learned at the largest companies in the world is that business is the business of selling happiness. Whatever you build, if you can frame the mission as how this increases happiness, the money and the customers follow. She closed by pointing out that Korean success was built on communities that helped each other farm terraced rice fields, and that you only succeed when you see the founder beside you as a colleague growing in the same direction rather than a rival competing for the same capital.

What companies should take from this

Put the evening into the language of daily work and four things are left.

1. Redesign on the assumption that knowledge work gets automated. Work that processes data, analyzes it and issues recommendations can be replaced by AI within 2 to 5 years, so organizations and individuals alike should treat pivoting and redefinition as a constant.
2. Open the window on the United States now. The window for entering the country and raising money narrows within 1 to 2 years, so secure a co-founder who knows the local culture and a community to stand in before anything else.
3. Choose the hub that fits the purpose. SaaS belongs in Silicon Valley and deep tech, physical AI and infrastructure belong somewhere like New Mexico, so pick the base of capital and talent that matches the character of the business.
4. Frame the mission as happiness. Capital and customers follow only when the product can explain how it increases happiness, and scale opens up when you grow alongside a community.

Frequently asked questions (FAQ)

Q: Is AI really going to replace knowledge work?

• Most work that takes in data, processes and analyzes it and returns a recommendation is in scope. The occupational research from OpenAI and Anthropic finds that the top 40 to 60 occupations can already be performed by AI, and the diagnosis that night was that the need for knowledge workers could shrink sharply within 2 to 5 years.

Q: When should we start on the United States and on fundraising?

• Now. The window is open but may narrow within 1 to 2 years, so she said the condition for success is to get inside a community and secure a friend and a co-founder who understand the local culture first.

Q: Is there a US hub worth looking at other than Silicon Valley?

• Yes. Most Silicon Valley capital goes toward B2B SaaS, so deep tech such as robotics, physical AI, energy infrastructure or logistics may find a better match in Albuquerque, New Mexico, with its national laboratories and its sovereign wealth fund.

Q: If we cannot raise VC money, has the business failed?

• No. A unicorn and a lifestyle business are different things. VCs look for a company that can grow to a billion dollars across 100 countries, so if your company is not that target, missing out on investment does not mean failure.

Q: How do we set a good mission?

• Frame it as how this increases happiness. What the largest companies in the world had in common was that they sold happiness, and her conclusion was that money, customers and scale follow once that framing holds.

Watch it back

The video from this event is in production. Subscribe to the YouTube channel and you will get a notification when it goes up.

Join Bloom

Bloom builds offline rooms where people and technology meet. We run them in Seoul, and now beyond it.

• 💬 Discord community, where events are announced first
• ▶️ YouTube, full talks from past events
• 📸 Instagram, photos from the floor
• 🎟️ Upcoming events